CVE-2025-68131 is a vulnerability in the cbor2 library where reusing a CBORDecoder instance across multiple decode operations causes values marked with the shareable tag (28) to persist in memory. This persistence allows an attacker-controlled CBOR message to read sensitive data from previously decoded messages if the decoder is reused across trust boundaries. This affects Red Hat AI Inference Server, Red Hat Enterprise Linux AI, and Red Hat OpenShift AI when processing untrusted CBOR data with reused decoders. The vulnerability is rated moderate by Red Hat and has a CVSS base score of 5.3 (Red Hat) and 7.5 (NVD). The vulnerability is related to CWE-212 (Improper Removal of Sensitive Information Before Storage or Transfer).

An attacker who can supply CBOR-encoded data to a vulnerable Red Hat AI Inference Server instance that reuses CBORDecoder objects can potentially read sensitive data from previously decoded messages. This leads to information disclosure across trust boundaries. The impact is limited to confidentiality with no integrity or availability effects reported.

No official patch or fix is currently available for this vulnerability. Red Hat advises that applications using the cbor2 library should avoid reusing CBORDecoder instances when processing data from different trust levels. If reuse is unavoidable, ensure that sensitive data is not processed by a decoder instance that will later handle untrusted input. This operational control prevents information leakage. Monitor Red Hat advisories for updates on fixes.